Here is the procedure for a complete recovery of the server
after a disaster has destroyed it. In this example virtual volumes
are not used. The steps are performed by the onsite administrator
unless otherwise indicated.
Procedure
Complete the following steps.
Restriction: When
you run the disaster recovery script or the commands that the script
contains, the determination must be made whether to run as root or
as the DB2® instance user ID.
- Review the recovery steps that are described in the RECOVERY.INSTRUCTIONS.GENERAL
stanza of the plan.
- Request the server backup tapes from the offsite vault.
- Break out the recovery plan file stanzas into multiple
files (see Breaking out a disaster recovery plan file.)
These files can be viewed, updated, printed, or run as Tivoli® Storage
Manager macros or
scripts.
- Print the RECOVERY.VOLUMES.REQUIRED file. Give the printout
to the courier to retrieve the copy storage pool volumes and active-data
pool volumes.
- Find a replacement server. The RECOVERY.DEVICES.REQUIRED
stanza specifies the device type that is needed to read the backups.
The SERVER.REQUIREMENTS stanza specifies the disk space required.
Note: When you use the Tivoli Storage
Manager device driver
(ADSMSCSI), you must start ADSMSCSI.
- The recovery media names and their locations are specified
in the RECOVERY.INSTRUCTIONS.INSTALL stanza and the MACHINE.RECOVERY.MEDIA.REQUIRED
stanza. Ensure that the environment is the same as when the disaster
recovery plan file was created. The environment includes
the following directory structures:
- The directory structure of the Tivoli Storage
Manager server executable
and disk formatting utility
- The directory structure for Tivoli Storage
Manager server configuration
files (disk log, volume history file, device configuration file, and
server options file)
- The directory structure and the files that are created when the
disaster recovery plan file was split into multiple files
- Restore the operating system and the Tivoli Storage
Manager server software
to the replacement server in one of the following ways:
- Build a new replacement server instead of restoring the environment
from a backup:
- Install the Tivoli Storage
Manager server
software
- Create the database instance user ID and group as in the original.
- Create the database directories, the active directories, and the
archive directories as in the original.
- Run the dsmicfgx utility to configure the replacement
instance. This step configures the API for the DSMSERV RESTORE DB
utility.
- Specify the instance user ID and password.
- Specify the database directories, the active directories, and
the archive directories.
- Remove the database instance that was created by the dsmicfgx utility.
For example,
dsmserv removedb TSMDB1
- Restore the original dsmserv.opt, volume history, and device configuration
files to the instance directory (as done by the recovery script in
plan file).
- Run the DSMSERV RESTORE DB (as done by the recovery script in
the plan file).
- Use a mksysb or Sysback tape. This tape, which includes the Tivoli Storage
Manager server software,
is created whenever software updates or configuration changes are
made to the AIX® system. The
tape location might be specified in the RECOVERY.INSTRUCTIONS.INSTALL
stanza.
Restoration from the mksysb tapes includes re-creating the
following items:
- The root volume group
- The file system where the database, active log, and archive log
directories, and the storage pool and disk volumes are located
After you use a mksysb or Sysback tape, issue the following
commands:
RUN DB2ILIST
RUN DB2 LIST DB DIRECTORY
Verify that the DB2 instance
and database information are restored. Create any additional AIX file
systems that might have contained directories that are needed by the
Tivoli Storage
Manager server.
Create the directories and set the ownership and group to match the
instance ownership. Verify the following file system directory types:
- Archive
- Active
- Failover
- Database
You can find these directories by issuing the
DB2 GET
DB CFG SHOW DETAIL command and reviewing the output or looking
in the recovery plan file.
- Restore the operating system
and the Tivoli Storage
Manager server
software to the replacement server in one of the following ways:
- Use a system restore utility that is appropriate for your
platform.
- Build a new replacement server instead of restoring the environment
from a backup:
- Install the Tivoli Storage
Manager server
software
- Create the database instance user ID and group as in the original.
- Create the database directories, the active directories, and the
archive directories as in the original.
- Run the dsmicfgx utility to configure the replacement
instance. This step configures the API for the DSMSERV RESTORE DB
utility.
- Specify the instance userid and password.
- Specify the database directories, the active directories, and
the archive directories.
- Remove the database instance that was created by the dsmicfgx utility.
For example
dsmserv removedb TSMDB1
- Restore the original dsmserv.opt, volume history, and device configuration
files to the instance directory (as done by the recovery script in
plan file)
- Run the DSMSERV RESTORE DB (as done by the recovery script in
the plan file.
- Restore the operating system, the Tivoli Storage
Manager server software,
the Tivoli Storage
Manager licenses,
and the administrative client on the replacement hardware.
- Build a new replacement server instead of restoring
the environment from a backup:
- Install the Tivoli Storage
Manager server
software.
- Create the database directories, the active directories, and the
archive directories as in the original.
- Run the dsmicfgx utility to configure the replacement
instance. This step configures the API for the DSMSERV RESTORE DB
utility.
- Specify the instance userid and password.
- Specify the database directories, the active directories, and
the archive directories.
- Remove the database instance that was created by the dsmicfgx utility.
For example,
dsmserv removedb TSMDB1
- Restore the original dsmserv.opt, volume history, and device configuration
files to the instance directory (as done by the recovery script in
plan file)
- Run the DSMSERV RESTORE DB (as done by the recovery script in
the plan file).
- Review the Tivoli Storage
Manager macros that
are contained in the recovery plan:
- If, at the time of the disaster, the courier had not picked up
the previous night's database and storage pool incremental backup
volumes but they were not destroyed, remove the entry for the storage
pool backup volumes from the COPYSTGPOOL.VOLUMES.DESTROYED file.
- If, at the time of the disaster, the courier had not picked up
the previous night's database and active-data pool volumes but
they were not destroyed, remove the entry for the active-data pool
volumes from the ACTIVEDATASTGPOOL.VOLUMES.DESTROYED file.
- If some required storage pool backup volumes could not
be retrieved from the vault, remove the volume entries from the COPYSTGPOOL.VOLUMES.AVAILABLE
file.
If some required active-data pool volumes could
not be retrieved from the vault, remove the volume entries from the
ACTIVEDATASTGPOOL.VOLUMES.AVAILABLE file.
- If all primary volumes were destroyed, no changes are required
to the PRIMARY.VOLUMES script and Tivoli Storage
Manager macro files.
- Review the device configuration file to ensure that the
hardware configuration at the recovery site is the same as the original
site. Any differences must be updated in the device configuration
file. Examples of configuration changes that require updates to the
configuration information are:
- Different device names
- Use of a manual library instead of an automated library
- For automated libraries, the requirement of manually placing the
database backup volumes in the automated library and updating the
configuration information to identify the element within the library.
This allows the server to locate the required database backup volumes.
For information about updating the device configuration file,
see Updating the device configuration file.
- To restore the database to a point where clients can be
recovered, ensure that the Tivoli Storage
Manager server is
halted and then invoke the RECOVERY.SCRIPT.DISASTER.RECOVERY.MODE
script file. Enter the script file name at the command prompt. As
an alternative, you can use the recovery script as a guide and manually
issue the steps.
The following steps are an example
recovery script:
- Copy the Tivoli Storage
Manager server
options file from the dsmserv.opt file to its original location.
- Copy the volume history file that is required by database restore
processing from the VOLUME.HISTORY.FILE file to its original location.
Use this copy of the volume history file unless you have a more recent
copy (after the disaster occurred).
- Copy the device configuration file that is required by database
restore processing from the DEVICE.CONFIGURATION.FILE file to its
original location.
- Issue the DSMSERV RESTORE DB command.
- Start the server.
- Register Tivoli Storage
Manager server
licenses.
- Mark copy storage pool volumes and active-data pool volumes retrieved
from the vault as available.
- Mark copy storage pool volumes and active-data pool volumes that
cannot be obtained as unavailable.
- Mark primary storage pool volumes as destroyed.
Due to
changes in hardware configuration during recovery, you might have
to update the device configuration file in the restored Tivoli Storage
Manager database
(see Updating the device configuration file
You
can mount copy storage pool volumes and active-data pool volumes upon
request, check in the volumes in advance, or manually place the volumes
in the library and ensure consistency by issuing the AUDIT
LIBRARY command. Use the AUDIT LIBRARY command to ensure
that the restored Tivoli Storage
Manager database
is consistent with the automated library volumes.
- If client workstations are not damaged, invoke the RECOVERY.SCRIPT.NORMAL.MODE
script file to restore the server primary storage pools. If client
workstations are damaged, you may want to delay this action until
after all clients are recovered.
This action is optional
because Tivoli Storage
Manager can
access the copy storage pool volumes and active-data pool volumes
directly to restore client data. Using this feature, you can minimize
client recovery time because server primary storage pools do not have
to be restored first. However, in this scenario, the client workstations
were not damaged, so the focus of the administrator is to restore
full Tivoli Storage
Manager server
operation.
As an alternative, you can use the recovery script
as a guide and manually run each step. The steps run in this script
are:
- Create replacement primary volumes.
- Define the replacement primary volumes to Tivoli Storage
Manager.
- Restore the primary storage pools from copy storage pools.
Attention: Restoring a primary storage pool from an active-data
pool might cause some or all inactive files to be deleted from the
database if the server determines that an inactive file must be replaced
but cannot find it in the active-data pool.
- Collect the database backup, copy storage pool volumes,
and active-data pool volumes used in the recovery for return to the
vault. For these backup volumes to be returned to the vault using
the routine MOVE DRMEDIA process, issue the following commands:
update volhist TPBK50 devcl=lib8mm ormstate=mountable
update volhist TPBK51 devcl=lib8mm ormstate=mountable
The
copy storage pool volumes and active-data pool volumes used in the
recovery already have the correct ORMSTATE.
- Issue the BACKUP DB command to back up the newly restored
database.
- Issue the following command to check the volumes out of
the library:
move drmedia * wherestate=mountable
- Create a list of the volumes to be given to the courier:
query drmedia * wherestate=notmountable
- Give the volumes to the courier and issue the following
command:
move drmedia * wherestate=notmountable
- Issue the PREPARE command.